How Do Macroeconomic and Bank-specific Variables Influence Profitability in the Austrian Banking Sector? Evidence from a Panel Vector Autoregression Analysis

What is it about?

Unlike most papers on bank profitability that only look at interest income, total income or any related ratios (e.g. ROA, ROE), we analyze the determinants of net interest income divided by total assets (NIM) and net fee and commission income divided by total assets (NFCIR) in a panel vector autoregression model where both variables are treated as endogenous. We describe a conceptual framework for the profit optimization problem faced by banks as a Bertrand game with differentiated products and intrafirm product interactions. In a second step, we empirically assess the implications of our conceptual framework using a unique supervisory data set of around 48,000 observations between 1998 and 2014. Apart from quantifying the contributions of the determinants (e.g. risk weighted assets, leverage ratio, loan loss provision ratio) to NIM and NFCIR, the empirical results show that interest income and fee and commission income should be regarded as strategic complements within a bank. In a third step, we add a second category of non-interest income as a predetermined variable, namely income from investment to the PVAR model. In contrast to NFCIR, income from investment can be regarded as a strategic substitute to the NIM within a bank. Finally, in a third set of results, we provide empirical evidence that portfolio separation between different loan and deposit categories (based on interest rates) does not hold in the Austrian banking system.

Featured Image

Why is it important?

(1) It is the first application of the newly implemented PVAR GMM estimator of Sigmund et al. (2016). (2) We treat different income categories of banks such as interest income and non-interest income as endogenous. (3) We also show that different deposit and loan rates are endogenous.